Remix.run Logo
▲ rolfus 14 hours ago

The technology is amazing, for sure. But the success of the current CI tech is also a testament to how fantastically adaptable the human brain is.

A normal cochlea has 15-20.000 tonotopically mapped hair cells that turn mechanical movement into electric signals which the brain then turn into meaningful sound. The fact that the brain can also make sense of the electrical signals coming out of just twelve to sixteen little electrical contacts placed along a silicon noodle that gets shoved into the cochlea, and turn this crude signal into into speech and music, is nothing short of mind blowing!

▲codedokode 11 hours ago | parent [-]

Should not it sound like a vocoder (which makes robotic voices)? I doubt one can hear any music this way though. Why they do not use more electrodes then?

▲jerf 20 minutes ago | parent | next [-]

Poking around for "cochlear implant simulator" online turns up a lot of stuff. An old YouTube video with various channel rates for speech and music: https://www.youtube.com/watch?v=SpKKYBkJ9Hw A report from someone with the implant on one side and a normal ear for the other: https://www.youtube.com/watch?v=1dhTWVMcpC4 What sounds like a personal report from someone who has had one for long enough for the brain to adjust to the input coming in and correctly amplify it: https://cochlearimplanthelp.com/journey/needing-a-cochlear-i...

▲retrac 2 hours ago | parent | prev [-]

A vocoder is a good approximation for what a cochlear implant's speech processor does.

There's been one person who was implanted, who had their natural hearing return for unknown reasons. So they can hear both cochlear, and quite well naturally in the same ear.

This let researchers at Arizona State University ask basically: what sounds most like the cochlear input? The various samples involving their research about 15 years ago are probably your best bet for what it sounds like.

Some people with musical training before retain the ability to hear arbitrary melodies with perfect pitch. Some do not.

There is basically no polyphony.